Perinatal Exposure to Bisphenol a Exacerbates Nonalcoholic Steatohepatitis-Like Phenotype in Male Rat Offspring Fed on a High-Fat Diet

JWEI, X SUN and others Maternal BPA exposure and 222:3 313–325 Research hepatic steatosis

Perinatal exposure to bisphenol A exacerbates nonalcoholic steatohepatitis-like phenotype in male rat offspring fed on a high-fat diet

Jie Wei2,*, Xia Sun1,*, Yajie Chen1, Yuanyuan Li, Liqiong Song, Zhao Zhou, Bing Xu, Yi Lin1 and Shunqing Xu

Key Laboratory of Environment and Health, Ministry of Education and Ministry of Environmental Protection, and State Key Laboratory of Environmental Health (Incubating), School of Public Health, Tongji Medical College, Correspondence Huazhong University of Science and Technology, Wuhan 430030, China should be addressed 1Key Laboratory of Urban Environment and Health, Department of Environmental and Molecular Toxicology, to S Xu or Y Lin Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China Emails 2Department of Basic Medical Sciences, Medical College, Xiamen University, Xiamen 361102, China [email protected] or *(J Wei and X Sun contributed equally to this work) [email protected]

Abstract

Bisphenol A (BPA) is one of the environmental endocrine disrupting chemicals, which is Key Words present ubiquitously in daily life. Accumulating evidence indicates that exposure to BPA " bisphenol A

Journal of Endocrinology contributes to metabolic syndrome. In this study, we examined whether perinatal exposure " steatosis to BPA predisposed offspring to fatty liver disease: the hepatic manifestation of metabolic " insulin resistance syndrome. Wistar rats were exposed to 50 mg/kg per day BPA or corn oil throughout " oxidative stress gestation and lactation by oral gavage. Offspring were fed a standard chow diet (SD) or a " inflammation high-fat diet (HFD) after weaning. Effects of BPA were assessed by examination of hepatic " fibrosis morphology, biochemical analysis, and the hepatic expression of genes and/or proteins involved in lipogenesis, fatty acid oxidation, gluconeogenesis, insulin signaling, inflammation, and fibrosis. On a SD, the offspring of rats exposed to BPA exhibited moderate hepatic steatosis and altered expression of insulin signaling elements in the liver, but with normal liver function. On a HFD, the offspring of rats exposed to BPA showed a nonalcoholic steatohepatitis-like phenotype, characterized by extensive accumulation of lipids, large lipid droplets, profound ballooning degeneration, impaired liver function, increased inflammation, and even mild fibrosis in the liver. Perinatal exposure to BPA worsened the hepatic damage caused by the HFD in the rat offspring. The additive effects of BPA correlated with higher levels of hepatic oxidative stress. Collectively, exposure to BPA may be a new risk factor for the development of fatty liver disease and further studies should assess whether

this ﬁnding is also relevant to the human population. Journal of Endocrinology (2014) 222, 313–325

http://joe.endocrinology-journals.org Ñ 2014 Society for Endocrinology Published by Bioscientiﬁca Ltd. DOI: 10.1530/JOE-14-0356 Printed in Great Britain Downloaded from Bioscientifica.com at 09/25/2021 11:58:55PM via free access Research JWEI, X SUN and others Maternal BPA exposure and 222:3 314 hepatic steatosis

Introduction

Nonalcoholic fatty liver disease (NAFLD) is one of the free fatty acid (FFA) composition as well as upregulation of typical hepatic manifestations of metabolic syndrome genes involved in lipid biosynthesis (Marmugi et al. 2012). characterized by hepatic fat accumulation in the absence Most importantly, this study revealed a nonmonotonic of excess alcohol consumption. NAFLD affects 15–40% of dose–response curve in which lower doses of BPA (doses the general population in Western countries and 9–40% of within one order of magnitude of approximately 50 mg/kg people in Asian countries (Farrell & Larter 2006). A subset per day) were more effective than the higher dose of patients with NAFLD may progress to nonalcoholic (5000 mg/kg per day) for altering lipid biosynthesis steatohepatitis (NASH), a more severe form of hepatic pathways. damage associated with inflammation, fibrosis, cirrhosis, Exposure to an adverse environment during critical or hepatocellular carcinoma (Farrell & Larter 2006). In windows of development will affect the programing of addition to genetic and diet-related factors, exposure to metabolically active tissues in offspring, predisposing environmental pollutants has been shown to negatively individuals to metabolic disorders later in life. For example, affect key pathogenic events implicated in NAFLD. For maternal nutrient excess or restriction has been confirmed instance, chronic inhalation of low levels of ambient air to increase the risk of NAFLD in the offspring (McCurdy particulate matter (PM 2.5) has been was reported to et al. 2009, Bayol et al. 2010, Oben et al. 2010). For BPA, induce hepatic steatosis and lipid peroxidation and, results from several studies using animal models indicated moreover, increase hepatic inflammatory and fibrosis that metabolic disorders were observed in offspring when stage in mice (Tomaru et al. 2007, Tan et al. 2009). they were exposed to BPA during the critical period of Exposure to nitrosamines and nicotine has also been development. Perinatal exposure to BPA altered early found to play critical roles in the pathogenesis of NAFLD adipogenesis and increased the body weight of offspring (Tong et al. 2009, Azzalini et al. 2010, Friedman et al. 2012). (Rubin & Soto 2009, Somm et al. 2009). A brief exposure to Bisphenol A (BPA) is a well-known environmental BPA during pregnancy has also been found to initiate endocrine disrupting chemical, which is used extensively insulin resistance in adult offspring (Alonso-Magdalena in the manufacture of products containing epoxy resins et al. 2010). Consistent with these findings, results from our and polycarbonate, such as toys, food and beverage previous study have confirmed that perinatal exposure to containers, paper currency, etc. (Diamanti-Kandarakis BPA resulted in a higher body weight and body fat Journal of Endocrinology et al. 2009). Humans can be exposed to BPA present in percentage, a greater mass of white adipocytes, hyperlipi- commonly used products. Recently, growing evidence has demia, hyperleptinemia, and insulin intolerance in adult indicated that exposure to BPA is associated with an rat offspring (Wei et al. 2011). All of these detrimental increased risk of some metabolic diseases such as obesity, effects of BPA might increase the risk of NAFLD. insulin resistance, type 2 diabetes, dyslipidemia, etc. In this study, we employed a developmental BPA (Alonso-Magdalena et al. 2011, Shankar et al. 2012, Vom exposure paradigm in rats to test the hypothesis that Saal et al. 2012). The presence of these metabolic disorder perinatal exposure to BPA caused a predisposition to events has been documented to predict the future hepatic steatosis. Our findings supported this hypothesis development of NAFLD (Fan et al. 2005, Hamaguchi and further indicated that perinatal exposure to BPA et al. 2005). Indeed, a recent epidemiological study has coupled with the postweaning high-fat diet (HFD) related BPA exposure to abnormal concentrations of the significantly initiated hepatic oxidative stress, thereby liver enzymes g-glutamyltransferase (gGT) and alkaline predisposing offspring to hepatic inflammation and early phosphatase (ALP), two of the sensitive markers for fibrosis in adulthood. hepatocellular damage (Lang et al. 2008). In addition, K K low concentrations of BPA (10 4–10 12 M) have been reported to induce lipid accumulation and steatosis in Materials and methods hepatic HepG2 cells by disturbing mitochondrial function Ethics and releasing proinflammatory cytokines (Huc et al. 2012). An animal study also confirmed that oral exposure of adult Allanimalexperimentswere carried out humanely male CD1 mice to BPA (0, 5, 50, 500, and 5000 mg/kg per following the guidelines for the care and use of animals day) resulted in the accumulation of hepatic triglycerides established by Tongji Medical College and were approved (TGs) and cholesteryl esters (CHOL), changes in hepatic by the Ethics Committee of Tongji Medical College.

Animals (CAT) and glutathione S-transferase (GST)) in hepatic tissue homogenates were detected using the correspond- Wistar rats were housed in polypropylene cages with ing assay kits (Nanjing Jiancheng Bioengineering Insti- free access to food and water under standard conditions tute, Nanjing, China). (22G2 8C, 12 h light:12 h darkness cycles). Glass water bottles were used to avoid potential contamination from sources other than administration. Pregnant female rats Liver function tests m were administered corn oil or 50 g/kg per day BPA by oral At the age of 27 weeks, blood was collected from the orbital gavage from gestational day 0 to postnatal day 21. Offspring sinus in rat offspring for serum measurements following were culled to five males and five females in each litter after 16 h of fasting. The liver biomarkers, namely alanine delivery and were nursed with their own mother. After aminotransferase (ALT), aspartate aminotransferase (AST), weaning, offspring were supplied with a standard chow diet ALP, and gGT, were assayed using commercially available (SD) or a HFD and allowed to feed ad libitum.TheSD kits (Mindray) on a Hitachi 747 autoanalyzer. In addition, contained 20.5% crude protein, 4.62% crude fat, and 52.5% the serum TNF and IL6 levels were measured using the rat nitrogen-free extract (total calories 3.45 kcal/g, 12.05% Quantikine ELISA kit (R&D Systems, Minneapolis, MN, calories in fat). The HFD contained 23.15% crude protein, USA) according to the manufacturer’s instructions. 13.14% crude fat, and 50.94% nitrogen-free extract (total calories 4.15 kcal/g, 28.53% calories in fat). Both standard Homeostasis Model Assessment of Insulin Resistance chow and HFDs were purchased from Shanghai Slac Laboratory Animal Co., Ltd. (Shanghai, China). All experi- Homeostasis Model Assessment of Insulin Resistance ments were carried out on male offspring from different (HOMA-IR) and quantitative insulin sensitivity check litters in each treatment group. index (QUICKI) were calculated from fasting blood glucose and serum insulin values. HOMA-IR was calculated as ! Sample processing and morphology examination fasting glucose (mM/l) fasting insulin (mU/ml)/22.5 (Matthews et al. 1985, Bonora et al. 2000). QUICKI was At the age of 27 weeks, rat offspring were killed and livers calculated as 1/(ln (fasting insulin, (mU/ml)Cln (fasting were quickly dissected out and weighed. Portions of the glucose, mg/dl)) (Katz et al. 2000, Chen et al. 2005). liver lobes were fixed in 4% paraformaldehyde, embedded Journal of Endocrinology in paraffin, and stained with hematoxylin/eosin (H&E) for RNA preparation and mRNA quantification histological analyses. Liver sections were also stained by Masson’s trichrome to visualize hepatic collagen depo- Total RNA was extracted from frozen liver samples using sition. The remaining liver samples were frozen in liquid TRIzol reagent (Invitrogen) and was reverse transcribed nitrogen and stored at K80 8C for subsequent lipid assays using a RevertAid First Strand cDNA Synthesis Kit (Thermo and molecular analysis. Scientific, Rockford, IL, USA) according to the manuals. cDNA was amplified by real-time quantitative PCR using the Hepatic lipid and oxidative stress index assays SYBR Green PCR Master Mix (Applied Biosystems) on an ABI Prism 7900 sequence detection system (Applied Bio- Hepatic lipids were extracted according to Folch’s systems). Relative mRNA expression levels were determined K methods (Folch et al. 1957). Briefly, 200 mg of hepatic by the 2 DDCt method with b-actin as an internal reference. tissues were homogenized in PBS, and then were extracted The primer sequences are given in Table 1. with 2:1 (vol/vol) chloroform/methanol by shaking overnight at 4 8C. The fractions of the organic phase were Protein extraction and western blot analysis isolated, evaporated under nitrogen, resuspended in chloroform, and then were assayed for TGs, CHOL, FFAs, Protein was extracted from liver using lysis buffer and phospholipids (PPLs) using commercially available containing 50 mM Tris–HCl (pH 7.4), 150 mM NaCl, kits (Mindray, Shenzhen, China) on a Hitachi 747 0.1% SDS, 1 mM EDTA, 1% Triton X-100, 1 mM phenyl- autoanalyzer. Oxidative stress markers including malon- methylsulfonyl fluoride, and cocktails of protease and dialdehyde (MDA), glutathione (GSH), oxidized gluta- phosphatase inhibitors. Protein concentration was thione (GSSG), and antioxidant enzymes (superoxide determined using the BCA Protein Assay kit (Thermo dismutase (SOD), glutathione peroxidase (GPx), catalase Scientific, Rockford, IL, USA). For insulin signaling

Table 1 Primer sequences for quantitative real-time PCR

Gene names Size (bp) 50-Primer 30-Primer Accession number

Acta2 146 CGAAGCGCAGAGCAAGAGAGGG GCAAGGTCGGATGCTCCTCTGG NM_031004.2 Actb 153 CAACGGCTCCGGCATGTGC CTCTTGCTCTGGGCCTCG NM_031144.3 Cd36 128 CTCTGACATTTGCAGGTCCA AGTGGTTGTCTGGGTTCTGG NM_031561.2 Col1a1 88 CCTGCTGGTCCTAAGGGAGAGC CTCTCTCACCAGGCAGACCTCG NM_053304.1 Cpt1a 101 AAGGTGCTGCTCTCCTACCA CGACCTGAGAGGACCTTGAC NM_031559.2 Fasn 139 AGCCCAGGAACAACTCATCC ACACAGGGACCGAGTAATGC NM_017332.1 G6pc 99 GGGCATCAATCTCCTCTGGG TCCAGGACCCACCAATACGG NM_013098.2 Cybb 159 CGTGGAGTGGTGTGTGAATGCC GCCAACCGAGTCACAGCCACAT NM_023965.1 Il6 133 CCCAACTTCCAATGCTCTCC GGTTTGCCGAGTAGACCTCA NM_012589.1 Cyba 91 CAGATCGAGTGGGCCATGTGGG AGCGTCCCGCAGTAGCCACGAT NM_024160.1 Ncf1 149 CCCTGCATCCTATTTGGAGCCC CCTCGATGGCTTCACCCTCAGA NM_053734.2 PckI 80 CGGTGGGAACTCACTGCTTG CCAGCCACCCTTCCTCCTTA NM_198780.3 Ppara 122 GGGATCTTAGAGGCGAGCCA GGGCCACAGAGCACCAATCT NM_013196.1 Pparg 80 CTGACCCAATGGTTGCTGATTAC GGACGCAGGCTCTACTTTGATC NM_001145366.1 Srebf1 97 CATCGACTACATCCGCTTCTTACA GTCTTTCAGTGATTTGCTTTTGTGA XM_213329.5 Tgfb1 97 GCTAATGGTGGACCGCAACA TGCTTCCCGAATGTCTGACG NM_021578.2 Tnf 100 CCATGGCCCAGACCCTCACA TCCAGCTGCTCCTCCGCTTG NM_012675.3

Acta2, smooth muscle a actin; Actb, b actin; Cd36, fatty acid translocase; Col1a1, collagen, type I, a 1; Cpt1a, carnitine palmitoyltransferase 1a; Fasn, fatty acid synthase; Il6, interleukin 6; G6pc, glucose 6-phosphatase, catalytic subunit; Cybb, cytochrome b-245, b polypeptide; Cyba, cytochrome b-245, a polypeptide; Ncf1, neutrophil cytosolic factor 1; PckI, phosphoenolpyruvate carboxykinase 1 (soluble); Ppara, peroxisome proliferator-activated receptor a; Pparg, peroxisome proliferator-activated receptor g; Srebf1, sterol regulatory element-binding protein 1; Tgfb1, transforming growth factor, b 1; Tnf, tumor necrosis factor a.

experiments, rat offspring were fasted overnight, anesthe- (Abcam, ab66579, 1:1000); Collagen1a1 (COL1A1) (Santa tized with pentobarbital, and received intraperitoneal Cruz Biotechnology, Santa Cruz, CA, USA, sc-8784, injections of 0.75 U/kg biosynthetic human insulin 1:1000); SREBP1C (Santa Cruz Biotechnology, sc-8984, (Novolin R; Novo Nordisk AIS, Bagsvaerd, Denmark). 1:100); G6PC (Santa Cruz, sc-25840, 1:500); PEPCK (Santa After 5–10 min, the livers were harvested and the protein Cruz Biotechnology, sc-166778, 1:500); PPARA (Santa was prepared as described above. Cruz Biotechnology, sc-9000, 1:500); TGFB1 (Abcam, For western blots, equal amounts of protein (40 mg) ab92468, 1:2000), and ACTA2 (Abcam, ab5694, 1:2000). Journal of Endocrinology were resolved with SDS–PAGE and electrophoretically transferred onto PVDF membranes (Bio-Rad). The mem- Statistical analysis branes were blocked with 5% non-fat dehydrated milk and then were incubated with the primary antibody at 4 8C Data are expressed as meanGS.E.M. Statistical analysis was overnight. The following day, the membranes were performed using the SPSS 17.0 software (SPSS, Inc.). One- incubated with the HRP-conjugated secondary antibody way ANOVA followed by Bonferroni’s post hoc test was (Cell Signaling Technology, Beverly, MA, USA) at 1:2000 used for comparison of the relevant groups (Chan et al. dilutions. For immunodetection, membranes were 2013, Shen et al. 2014). A P value of !0.05 was considered visualized using enzyme-catalyzed chemiluminescence significant. (Amersham Biosciences). The density of the bands was quantified using the Image Pro Plus version 6.0 software Results (Media Cybernetics, Bethesda, MD, USA) and normalized Body and liver weight and hepatic lipid content to ACTB. The following primary antibodies were used: ACTB (Cell Signaling Technology, #4970, 1:2000); AKT1 As shown in Fig. 1A, the body weight was significantly (Cell Signaling Technology, #9272, 1:2000); Thr308- increased in offspring of rats exposed to BPA, regardless of phosphorylated AKT1 (Cell Signaling Technology, #9275, whether they were fed on a standard chow or HFD. There 1:500); GSK3B (Cell Signaling Technology, #9315, 1:2000); was no difference in either liver weight or calculated Ser9-phosphorylated GSK3B (Cell Signaling Technology, liver/body weight ratio between control offspring and #9336, 1:500); insulin receptor (INSR) (Cell Signaling offspring of rats exposed to BPA fed on a SD (Fig. 1A). Technology, #3025, 1:1000); Tyr1150/1151-phosphory- Compared with controls fed on a HFD, offspring of rats lated INSR (Cell Signaling Technology, #2969, 1:500); exposed to BPA fed on a HFD displayed 154 and 132% IL6 (Abcam, Cambridge, MA, USA, ab8339, 1:1000); TNF increases in the liver weight and the calculated liver/body

AB1000 25 6.0 Control 800 BPA *** 20 *** 4.8 ### ### ** 600 # 15 3.6

400 10 2.4 ## Body weight (g) 200 Liver TG (mg/g) 5 1.2

0 0 0.0 Standard High-fat Standard High-fat Standard High-fat 40 5 200 40 *** 32 *** 4 160 *** 32

mol/g) ## 24 3 µ 120 24 ### ### *** 16 2 80 ## 16 Liver PPL (mg/g) Liver CHOL (mg/g) Liver weight (g) 40 8 8 1 Liver FFA (

0 Liver/body weight ratio (%) 0 0 0 Standard High-fat Standard High-fat Standard High-fat Standard High-fat

C D 300 150 Control BPA Control 240 BPA 120 ** 180 90 *** 120 60 ALT (U/I) AST (U/I) # # 60 30

0 0 Standard Standard High-fat Standard High-fat 60 150

48 120 *** * ### 36 90

GT (U/I) 24 60 ALP (U/I) γ High-fat 12 30

0 0 Standard High-fat Standard High-fat

Figure 1 The effect of BPA on liver injury. (A) Body weight (wt), liver weight, stained with hematoxylin and eosin dye (400! magniﬁcation). Journal of Endocrinology and body liver percentage. (B) Hepatic lipid contents. Hepatic levels Arrowheads indicate lipid droplets and arrows indicate inﬂammatory of triglycerides (TGs), cholesterol (CHOL), free fatty acids (FFAs), and clusters. Data are expressed as meansGS.E.M.(nZ6 rats per group); only one phospholipids (PPLs) were measured. (C) Liver function tests. Serum level offspring was selected per litter. #P!0.05; ##P!0.01; ###P!0.001 relative of aspartate aminotransferase (AST), alanine aminotransferase (ALT), to the value for the control offspring fed on a SD. *P!0.05; **P!0.01; g-glutamyltransferase (gGT), and alkaline phosphatase (ALP) were ***P!0.001 relative to the value for the control offspring fed on a HFD. SD, measured. (D) Representative light microscopic micrographs of liver section standard chow diet; HFD, high-fat diet.

weight ratio respectively (Fig. 1A). For hepatic lipid Histological examination content, TGs and FFAs were increased in offspring of rats H&E-stained liver sections from offspring of control rats fed exposed to BPA compared with controls, regardless of on a SD revealed normal liver histology manifesting as whether they were fed on a standard chow or HFD normal cell size, with a prominent cell nucleus, uniform (Fig. 1B). Higher CHOL and lower PPLs were observed cytoplasm, and radially aligned hepatic cell cords, whereas only in offspring of rats exposed to BPA fed on a HFD, the offspring of rats exposed to BPA fed on the same diet compared with controls fed on a HFD (Fig. 1B). showed some features similar to those observed in human NAFLD, including disordered hepatic cell cords and Hepatic function test modestly increased lipid droplets (Fig. 1D). On a HFD, an On a SD, serum level of ALT was increased in offspring of rats evident injury was observed in the livers of the offspring of exposed to BPA compared with controls (Fig. 1C), but no both control rats and rats exposed to BPA including almost difference in AST, gGT, and ALP was observed between the complete absence of sinus hepaticus, disordered hepatic cell offspring of control rats and the offspring of rats exposed to cord, diffuse lipid droplets, and ballooning degeneration BPA. On a HFD, serum levels of ALT, AST, gGT, and ALP were (Fig. 1D). But notably, fat was observed mainly as higher in the offspring of rats exposed to BPA than in macrovesicular droplets in the offspring of rats exposed to controls fed on a HFD (Fig. 1C), indicative of liver injury. BPA fed on a HFD. Meanwhile, a higher extent of

ballooning and inflammatory clusters was also displayed in The mRNA expressions of Pparg and Cd36 were elevated the livers of the offspring of rats exposed to BPA compared in BPA-exposed offspring compared with the control under with controls fed on a HFD (Fig. 1D). high-fat dietary conditions (Fig. 2A). Although the levels of PPARA (Ppara) and one of its targets Cpt1a were compensatorily increased in the offspring of rats exposed Levels of genes related to energy metabolism in liver to BPA fed on a SD, both genes were decreased in the Genes coding for key factors involved in energy meta- offspring of rats exposed to BPA fed on a HFD compared bolism were examined in the offspring of rats exposed to with controls fed on a HFD (Fig. 2). Finally, two key BPA and control rats. As shown in Fig. 2, both mRNA and regulators of hepatic gluconeogenesis, PCK1 and G6PC, protein expressions of lipogenic transcription factor were examined. It can be observed from Fig. 2 that BPA had SREBP1C (Srebf1) were significantly increased in the livers no significant effect on expression of either PCK1 (Pck1)or of the offspring of rats exposed to BPA, regardless of G6PC (G6pc) in offspring fed on a SD. However, the mRNA whether rats are fed on a SD or a HFD. Similarly, mRNA and protein levels of both genes were increased in the expression of its target gene Fasn was coordinately offspring of rats exposed to BPA fed on a HFD compared upregulated in the offspring of rats exposed to BPA. with controls fed on a HFD.

A Lipogenesis 5 Srebf1 5 Fasn 8 Pparg 5 Cd36 4 4 *** 4 6 *** *** 3 3 3 ## 4 *** ## 2 # 2 # 2 2 1 1 1 Relative mRNA levels Relative mRNA levels Relative mRNA levels 0 Relative mRNA levels 0 0 0

Fatty acid oxidation Gluconeogenesis

2.0 Ppara 2.0 Cpt1a 4 G6pc4 Pck1 # ## 1.5 3 3 ** 1.5 ** ## 1.0 1.0 2 ## 2 Journal of Endocrinology 0.5 * 0.5 * 1 1 Relative mRNA levels Relative mRNA levels Relative mRNA levels 0.0 Relative mRNA levels 0.0 0 0

B 4 SREBP1C 2.0 PPARA *** BPA – + – + 3 1.5 HFD – – + + # SREBP1C 2 1.0 # ### ### PPARA 1 0.5 Relative protein levels Relative protein levels G6PC 0 0.0 PCK1 2.5 G6PC3.0 PCK1 ACTB 2.0 ** 2.4 *** 1.5 1.8

1.0 1.2 Control SD Control HFD BPA SD BPA HFD 0.5 0.6 Relative protein levels 0.0 Relative protein levels 0.0

Figure 2 The effect of BPA on hepatic energy metabolism in the liver. (A) mRNA three replicates for each gene; for protein, nZ3 rats per group, three expression of genes associated with lipogenesis, fatty acid oxidation, and replicates for each gene; only one offspring was selected per litter. gluconeogenesis. (B) Levels of key proteins involved in hepatic energy #P!0.05; ##P!0.01; ###P!0.001 relative to the control offspring fed on a metabolism. The expression levels were normalized to the ACTB internal SD. *P!0.05; **P!0.01; ***P!0.001 relative to the control offspring fed control and were expressed as fold of values in control offspring fed on a on a HFD. SD, standard chow diet; HFD, high-fat diet. SD. Data are expressed as meansGS.E.M. For mRNA, nZ5 rats per group,

Insulin signaling elements in the liver of INSR, AKT1, and GSK3B were similar between the offspring of control rats and the offspring of rats exposed Hepatic steatosis is strongly associated with insulin to BPA, regardless of whether they were fed on a SD or a resistance (Marchesini et al. 2005, Utzschneider & Kahn HFD. Meanwhile, p-INSR to INSR ratio, p-AKT1 to AKT1 2006), hence the HOMA-IR and the QUICKI were ratio, and p-GSK3B to GSK3B ratio were also significantly calculated. It is shown in Fig. 3A that the HOMA-IR was less in BPA-exposed offspring than in the control offspring higher, whereas the QUICKI was lower in the offspring of fed on the same diet (Fig. 3B). rats exposed to BPA compared with the offspring of control rats regardless of whether they were fed on a SD or Hepatic oxidative stress a HFD. Perinatal exposure to BPA decreased hepatic insulin sensitivity and induced insulin resistance in the Indexes related to oxidative stress were measured in rat offspring. As insulin resistance could result from the livers of rat offspring. On a SD, the offspring of rats defective insulin signaling in peripheral tissues, the exposed to BPA exhibited a lower GSH:GSSG ratio activation of insulin signal elements in the liver was compared with controls (Fig. 4A). On a HFD, perinatal determined by western blot in the rat offspring after exposure to BPA increased liver GSSG levels in the insulin injection. It is shown in Fig. 3B that the levels of offspring, with significant diminution in hepatic GSH insulin-stimulated p-INSR, p-AKT1, and p-GSK3B were and decreases in the GSH:GSSG ratio (Fig. 4A). The levels significantly downregulated in the livers of the offspring of of a lipid peroxidation product MDA, another biomarker rats exposed to BPA compared with controls, but the levels of oxidative stress, were significantly increased in the liver

A 150 60 20 0.25

120 48 16 *** 0.20 *** U/ml) Control SD ### µ 90 36 *** 12 0.15 ## BPA SD ## Control HFD ### ** 60 24 ### 8 (mg/dl) 0.10 BPA HFD QUICK HOMA-IR

Blood glucose 30 12 # 4 # 0.05

0 Serum insulin ( 0 0 0.00

1.5 B 0.4 1.2 BPA – + – + – + – + 0.3 # HFD ––+ + – – + + 0.9 # ### 0.2 ## Journal of Endocrinology Insulin – – ––+ + + + 0.6 *** *** 0.1

to value for actin) 0.3 p-INSR:INSR ratio p-INSR

p-INSR/INSR (fold relative 0.0 0.0 p-INSR Total INSR

INSR 1.5 1.2 1.2 0.9 0.9 ### p-AKT1 0.6 ### 0.6 ### ### 0.3 0.3 to value for actin) * * AKT1 p-AKT1:AKT1 ratio

p-AKT1/AKT1 (fold relative 0.0 0.0 p-AKT1 Total AKT1

p-GSK3B 1.5 1.6 ratio

1.2 1.2 ## 0.9 ### GSK3B 0.8 0.6 ### ### *** 0.4 0.3 *** to value for actin)

ACTB p-GSK3B:GSK3B 0.0 0.0

p-GSK3B/GSK3B (fold relative p-GSK3B Total GSK3B

Figure 3 The effect of BPA on hepatic insulin resistance. (A) Fasting blood glucose, Protein levels were quantiﬁed relative to a ACTB loading control and were serum insulin, and the calculation of homeostasis model assessment of expressed as fold of values in control offspring fed on a SD. Data are insulin resistance (HOMA-IR) and quantitative insulin sensitivity check expressed as meansGS.E.M. of three replicates for each protein (nZ3rats index (QUICKI) (nZ6 rats per group; only one offspring was selected per per group; only one offspring was selected per litter). #P!0.05, ##P!0.01, litter). (B) Regulation of the key insulin signaling element protein ###P!0.001 relative to the control offspring fed on a SD. *P!0.05, phosphorylation in the liver. Rat offspring were fasted overnight and **P!0.01, ***P!0.001 relative to the control offspring fed on a HFD. received intraperitoneal injections of saline or insulin (0.75 U/kg). Liver SD, standard chow diet; HFD, high-fat diet. tissues were harvested 10 min later for immunoblotting analysis.

A B 40 10 10 12 Control *** BPA 8 8 30 9 * # 6 6 # 20 ### 6 MDA GSH GSSG 4 4 **

GSH/GSSG 3 10 2 *** 2 (nmol/mg protein) (nmol/mg protein) (nmol/mg protein) 0 0 0 0 Standard High-fat Standard High-fat Standard High-fat Standard High-fat C 100 100 40 120 80 80 30 90 # 60 60 ## ** 20 60 40 40 ## * *** 10 20 20 *** 30 CAT (U/mg protein) GPx (U/mg protein) GST (U/mg protein) SOD (U/mg protein) 0 0 0 0 Standard High-fat Standard High-fat Standard High-fat Standard High-fat

D Cyba Ncf1 Cybb 8 12 6.0 *** 6 *** 9 *** 4.5 Control SD BPA SD 4 6 3.0 Control HFD BPA HFD 2 3 1.5 Relative mRNA levels Relative mRNA levels Relative mRNA levels 0 0 0.0

Figure 4 The effect of BPA on hepatic oxidative stress. (A) Glutathione (GSH) values in control offspring fed on a SD. Data are expressed as meansGS.E.M. homeostasis. (B) Levels of malondialdehyde (MDA). (C) Activities of (A, B, and C), nZ6 rats per group; (D), nZ5 rats per group, three replicates antioxidant enzymes, including catalase (CAT), glutathione peroxidase for each gene; only one offspring was selected per litter. #P!0.05, (GPx), glutathione S-transferase (GST), and superoxide dismutase (SOD). ##P!0.01, ###P!0.001 relative to the control offspring fed on a SD. (D) NADPH oxidase subunit mRNA expression. Expression levels were *P!0.05, **P!0.01, ***P!0.001 relative to the control offspring fed on a normalized to the Actb internal control and were expressed as fold of HFD. SD, standard chow diet; HFD, high-fat diet. Journal of Endocrinology

homogenates of the offspring of rats exposed to BPA controls fed on a HFD (Fig. 5A). Consistent with the compared with that measured in controls fed on a HFD increased peripheral blood levels of inflammatory (Fig. 4B). Meanwhile, the levels of the antioxidant mediators, exposure to BPA also significantly increased enzymes, CAT, GPx, GST, and SOD, were reduced in the the mRNA and protein levels of TNF and IL6 in the liver of offspring of rats exposed to BPA fed on a HFD compared offspring fed on a HFD (Fig. 5B). Using Masson’s trichrome with controls fed on a HFD, as shown in Fig. 4C. staining, results shown in Fig. 5C indicated that no Additionally, the mRNA expression of ROS-generating significant histological signs of hepatic fibrosis appeared NADPH oxidases, such as Cyba, Ncf1, and Cybb, were in offspring of both groups fed on the SD and the HFD. significantly upregulated in the offspring of rats exposed However, incipient perivenular fibrosis was observed in to BPA fed on a HFD (Fig. 4D). No significant difference the livers of the offspring of rats exposed to BPA fed on a was detected in the mRNA expression of these genes HFD (Fig. 5C). Consistent with staining results, both between the offspring of control rats and the offspring of mRNA and protein levels of key fibrotic markers, including rats exposed to BPA fed on a SD (Fig. 4D). ACTA2, COL1A1, and TGFB1, were increased in the livers of the offspring of rats exposed to BPA fed on a HFD compared with controls fed on a HFD (Fig. 5D). Inflammation and fibrosis

On a SD, no difference in the serum inﬂammatory Discussion cytokines IL6 and TNF was observed between the offspring of rats exposed to BPA and the offspring of control rats BPA is one of the most widely used industrial chemicals (Fig. 5A). But on a HFD, elevated levels of both IL6 and TNF with applications in the manufacture of plastic products were shown in BPA-exposed offspring compared with that can increase the risk of chronic metabolic disease.

AB 160 BPA – + – + 6 6 HFD ––+ + ** 5 *** 5 120 *** IL6 4 4 # *** 80 *** 3 3 *** IL6 # # TNF TNF 2 2 40 Pro-inflammation cytokines (pg/ml) 1 1 β-actin (relative expression) (relative 0 expression) (relative 0 0 IL6 TNF mRNA Protein mRNA Protein

C Control SDBPA SD Control HFD BPA HFD

D Control ND BPA ND BPA – + – + Control HFD HFD ––++ BPA HFD 5 5 5 ACTA2 4 4 4 *** ** *** COL1a1 3 3 3 * * *** ACTA2 2 2 TGFB1 2

TGFB1 COL1a1 1 1 1 (relative expression) (relative (relative expression) (relative ACTB 0 expression) (relative 0 0 mRNA Protein mRNA Protein mRNA Protein

Figure 5 The effect of BPA on hepatic inflammation and fibrosis response. (A) Serum on a SD. Data are expressed as meansGS.E.M. For mRNA, nZ5 rats per concentrations of inflammatory cytokines: IL6 and TNF (nZ6). (B) mRNA group, three replicates for each gene; for protein, nZ3 rats per group, and protein levels of IL6 and TNF. (C) Representative Masson’s trichrome three replicates for each gene; only one offspring was selected per litter. stains of livers (200! magnification). The collagen fibers are stained blue. #P!0.05 relative to the control offspring fed on a SD. *P!0.05, **P!0.01, (D) mRNA and protein levels of key fibrotic markers: ACTA2, COL1A1, ***P!0.001 relative to the control offspring fed on a HFD. SD, standard Journal of Endocrinology and TGFB1. The expression levels were normalized to the ACTB internal chow diet; HFD, high-fat diet. control and were expressed as fold of values in control offspring fed

With regard to this finding, we had previously shown that et al. 2011). Alterations in FA source, including circulating perinatal exposure to BPA resulted in multiple features FFAs from adipose tissue via lipolysis (59%), newly of metabolic syndrome in the offspring, including produced FAs within the liver through de novo lipogenesis increased body weight, hyperlipidemia, hyperleptinemia, (26%), and dietary FAs (15%) (Donnelly et al. 2005), would and insulin resistance (Wei et al. 2011). Considering that produce a fatty liver phenotype. In a pervious study, we had NAFLD is the typical hepatic manifestation of the observed that BPA exposure resulted in a higher body fat metabolic syndrome, the current study further investi- percentage and a significantly greater mass of adipocytes in gated the adverse effects of BPA on the liver. both offspring fed on a SD and those fed on a HFD. The In this study, perinatal exposure to BPA, even at the expansion of adipose tissue mediated by exposure to BPA current TDI of 50 mg/kg per day, induced abnormal made the organism release excess FFAs into the circulation accumulation of lipids in the liver of adult offspring. (Supplementary Figure 1, see section on supplementary These findings were in agreement with results described in a data given at the end of this article) and, moreover, previous paper demonstrating that male CD1 mice treated diminished the release of insulin-sensitizing and anti- with BPA at 50 mg/kgperdayfor28daysexhibited inflammatory cytokines and increased expression of accumulation of CHOLs and TGs in the liver (Marmugi proinflammatory molecules, finally resulting in fatty liver et al. 2012). Hepatic steatosis results from a disturbance in and insulin resistance. Also, SREBP1C and its downstream the balance among TG delivery, synthesis, export, and lipogenic targets (FASN, PPARG, and CD36) were stimu- oxidation. In the liver, TGs are assembled by coupling fatty lated in the livers of the offspring of rats exposed to BPA, acids (FAs) to a glycerol backbone via ester bonds (Cohen indicating that BPA increased lipogenesis via prompting

de novo synthesis of FAs. Specifically, a significantly higher These observations were in agreement with results degree of lipid accumulation, together with impaired liver reported in a previous paper in which the level of tyrosine function, widespread ballooning degeneration, moderate phosphorylation of the INSR was decreased in the livers of inflammatory cell infiltration, and mild fibrosis, was mice exposed to 100 mg/kg per day BPA for 8 days (Batista exhibited in the livers of the offspring of rats exposed to et al. 2012). However, insulin-induced AKT1 phosphoryl- BPA when they consumed a HFD immediately after ation was similar for BPA-exposed and control mice in that weaning. A study conducted on rats demonstrated that study, which may be due to a compensation mechanism dietary fats taken up in the intestine could be packaged that occurred at the early stage of insulin resistance, before into TG-rich chylomicrons and 20% of TGs in chylomi- the onset of hyperglycemia. In this study, BPA-exposed crons were then delivered to the liver (Redgrave 1970). offspring had developed a higher degree of insulin Extrapolating from these experiments, the HFD (20 g food resistance, characterized by hyperglycemia, hyperinsuli- per day) could furnish the liver with a significantly higher nemia, and glucose intolerance. We thus suggested that amount of fat (approximately 0.53 g fat each day) than BPA decreased expression of p-AKT1 and p-GSK3B in the the SD (approximately 0.18 g fat each day). Postweaning livers of offspring, thereby leading to an increase in the HFD intake was responsible, at least in part, for excess hepatic glucose output and finally contributing to the FA supply to the liver. continuous hyperglycemia. In addition, two gluconeo- In response to the large pool of FAs, genes involved in genic key enzymes, G6PC and PCK1, were elevated in the the FA b-oxidation pathway, such as Ppara and its target livers of the offspring of rats exposed to BPA fed on a HFD. gene Cpt1a, were altered by BPA in livers of offspring. Ppara Normally, insulin is able to rapidly and substantially is the predominant PPAR subtype in rodent liver, which inhibit G6PC and PCK1 and gluconeogenesis in the liver. plays a major role in controlling genes involved in FA However, the offspring of rats exposed to BPA fed on a HFD elongation, desaturation, oxidation, and transport (Reddy exhibited lack of suppression of these gene transcription 2001). Cpt1a is a membrane transporter of FAs from the processes during hyperinsulinemia. Glucose overproduc- cytoplasm into the mitochondrial matrix and is a primary tion would lead to further increases in insulin secretion, rate-limiting enzyme involved in FA oxidation (McGarry finally contributing to the vicious cycle that triggered the & Foster 1980). Although exposure to BPA caused liver to become highly insulin-resistant. unexpected activation of Ppara and Cpt1a in offspring Notably, AKT1 is known to be a major regulator of fed on a SD, it may be only a compensatory response to lipogenesis in the liver. AKT1 activates SREBP1C via AKT1- Journal of Endocrinology metabolize the excessive FAs in the liver. In contrast to mediated phosphorylation and inhibition of GSK3. In this these findings, the offspring of rats exposed to BPA fed on a study, exposure to BPA significantly impaired AKT1 HFD had a notable reduction in both Ppara and Cpt1a, signaling in the liver. Under these circumstances, it which was indicative of inhibition of FA oxidation and appeared that transcriptional activity of SREBP1C might whichinturnwouldfurtherleadtohepaticlipid be repressed by increased GSK3 activity, leading to the accumulation. BPA adversely influenced the coordinated reduced expression of downstream lipogenic genes. But modulation of hepatic uptake, synthesis, and oxidation of surprisingly, hepatic TGs and lipogenic gene expression FAs, thereby causing accumulation of toxic lipid-derived were increased after exposure to BPA even in the absence of metabolites in hepatocytes. hepatic insulin signaling, indicating that insulin resistance In this study, marked increases in HOMA-IR and (impaired activation of AKT1 pathway) and hypersensitiv- decreases in QUICKI were established in the offspring of ity to insulin (increased expression of SREBP1C) existed rats exposed to BPA, indicating that perinatal exposure to concomitantly in the livers of the offspring of rats exposed BPA induced insulin resistance in the rat offspring. The to BPA. These results were consistent with the observation liver is a key organ for glucose and lipid metabolism, hence that the expression of SREBP1C was increased, whereas we subsequently measured whether exposure to BPA the insulin-induced activation of AKT1 was inhibited in induced insulin resistance by blocking the hepatocellular the livers of ob/ob mice (Shimomura et al. 2000). We insulin signaling pathway. We found that autophosphor- interpreted these data to indicate that hyperinsulinemia ylation of the INSR on Tyr1150/1151 in response to was able to stimulate SREBP1C activity in the livers of the insulin was decreased in the livers of offspring after offspring of animals exposed to BPA through a pathway BPA treatment compared with controls, which was that did not require AKT1 activation. Further investi- followed by reduced AKT1 phosphorylation on Thr308 gations are still required to explore the signaling molecule and its downstream GSK3B phosphorylation on Ser9. involved in promoting SREBP1C activation.

According to the developmental origin hypothesis, we expression of DNA methyltransferase 3B mRNA in male rat suggested that exposure to BPA during critical stages of offspring at 3 and 21 weeks. Alterations in DNA methyl- development made the offspring develop responses to ation regulated the expression of certain important genes altered intrauterine environments, ultimately predisposing such as Gck, consequently contributing to insulin resist- them to fatty liver disease. Furthermore, developmental ance and impaired glucose intolerance in adulthood (Ma programing mediated by BPA affected postnatal defensive et al. 2013). In addition, exposure to BPA has also been responses to other stresses (such as HFD) in offspring. reported to interfere with DNA repair in the placental cell When the offspring of rats exposed to BPA were weaned line through alterations of cellular microRNA, further onto a HFD, several histopathological features reminiscent supporting the potential role of epigenetics in fetal of human NASH, such as micro- and macro-vesicular lipid reprograming by BPA-induced toxicity (Avissar-Whiting accumulation, ballooning, inflammation infiltration, the et al. 2010). Thus, we postulated that the BPA-induced gradual appearence of collagen fibrils, and impaired cell fatty liver phenotype may also be epigenetically associ- function, were detected in the liver. Aberrant epigenetic ated, although the way in which BPA affects the epigenetic modifications have been suggested to be one of the process is unclear and remains to be demonstrated in important biological mechanisms linking fetal exposures future studies. to metabolic disease progression. The epigenome can In this study, another mechanism of the liver control and influence gene expression profiles of most pathology of steatohepatitis observed in BPA-exposed organs and cell types during development, both in utero offspring upon a feeding with a high-fat diet was oxidative and throughout life, through DNA methylation, histone stress and its downstream events, as we identified modifications, and a variety of non-coding RNAs (Bernal & accumulation of MDA. Perinatal exposure to BPA induced Jirtle 2010). BPA has been reported to be epigenetically hepatic oxidative stress in offspring fed on a HFD, possibly toxic, which can induce aberrant epigenetic modifications via the upregulation of genes involved in the generation in early life (Singh & Li 2012, Veiga-Lopez et al. 2013). For of ROS (e.g. the NADPH oxidase family Cyba, Ncf2, and example, perinatal exposure to 50 mg/kg per day BPA Cybb) and the reduction in both enzymatic and non- decreased hepatic global DNA methylation and increased enzymatic antioxidants. ROS along with the products of

Maternal BPA exposure Journal of Endocrinology Altered intrauterine environments

‘First hit’

p-INSR p-AKT1 p-GSK3B Hepatic insulin resistance × Insulin ? SREBP1C FASN/CD36/PPARG Increased de novo lipogenesis

PPARA CPT1A Reduced FA oxidation

Nonalcoholic fatty liver phenotype

‘Second hit’ High-fat diet

Ongoing lipid accumulation Oxidative damage

Inflammation Fibrosis

Nonalcoholic steatohepatitis-like phenotype

Figure 6 Proposed pathways by which fatty liver disease is programed in BPA-exposed offspring.

lipid peroxidation could activate nuclear factor kB, which in animal experiments and data processing. Y C carried out additional in turn induced the release of several cytokines, including experiments in the revision of the manuscript. All authors have contributed to, seen, and approved the manuscript. TNF, IL6, TGFB1, etc., in the BPA-exposed offspring upon high-fat feeding. Upregulation of proinflammatory cytokines TNF and IL6 could induce hepatic inflammation Acknowledgements through delivering polymorphonuclear and mononuclear The authors thank the Animal Laboratory of Wuhan University for leukocytes into inflamed tissues, and increased TGFB1 was providing animal management and care. able to activate hepatic stellate cells into collagen- producing myofibroblastic cells, promoting collagen synthesis in the liver. TNF and lipid peroxidation products References could also inhibit the electron transport chain of the Alonso-Magdalena P, Vieira E, Soriano S, Menes L, Burks D, Quesada I & mitochondria and worsen the mitochondrial function Nadal A 2010 Bisphenol A exposure during pregnancy disrupts glucose homeostasis in mothers and adult male offspring. Environmental Health (Nagakawa et al. 2005). Mitochondrial damage in turn Perspectives 118 1243–1250. (doi:10.1289/ehp.1001993) would lead to secondary production of ROS and inhibition Alonso-Magdalena P, Quesada I & Nadal A 2011 Endocrine disruptors in of the b-oxidation of lipids (Santra et al. 2007), further the etiology of type 2 diabetes mellitus. Nature Reviews. Endocrinology 7 346–353. (doi:10.1038/nrendo.2011.56) increasing the progression of hepatic steatosis, inflam- Avissar-Whiting M, Veiga KR, Uhl KM, Maccani MA, Gagne LA, Moen EL & mation, and fibrosis and starting a vicious circle. Marsit CJ 2010 Bisphenol A exposure leads to specific microRNA In summary, origins of metabolic diseases may lie alterations in placental cells. Reproductive Toxicology 29 401–406. in early life, even in utero, experiences. Perinatal exposure to (doi:10.1016/j.reprotox.2010.04.004) Azzalini L, Ferrer E, Ramalho LN, Moreno M, Domıńguez M, Colmenero J, BPA led to an excess of lipid accumulation and insulin Peinado VI, Barbera` JA, Arroyo V & Gine`s P 2010 Cigarette smoking resistance in the livers of the rat offspring. More impor- exacerbates nonalcoholic fatty liver disease in obese rats. Hepatology 51 tantly, perinatal exposure to BPA also resulted in the liver 1567–1576. (doi:10.1002/hep.23516) Batista TM, Alonso-Magdalena P, Vieira E, Amaral MEC, Cederroth CR, being ‘primed’ for more severe steatosis, inflammation, and Nef S, Quesada I, Carneiro EM & Nadal A 2012 Short-term treatment fibrosis. The postweaning HFD triggered a steatohepatitis with bisphenol-A leads to metabolic abnormalities in adult male phenotype in the offspring of rats exposed to BPA via mice. PLoS ONE 7 e33814. (doi:10.1371/journal.pone.0033814) Bayol SA, Simbi BH, Fowkes RC & Stickland NC 2010 A maternal “junk aggravation of hepatic oxidative stress (Fig. 6). food” diet in pregnancy and lactation promotes nonalcoholic fatty liver disease in rat offspring. Endocrinology 151 1451–1461. (doi:10.1210/en.2009-1192) Journal of Endocrinology Bernal AJ & Jirtle RL 2010 Epigenomic disruption: the effects of early Supplementary data developmental exposures. Birth Defects Research. Part A, Clinical and This is linked to the online version of the paper at http://dx.doi.org/10.1530/ Molecular Teratology 88 938–944. (doi:10.1002/bdra.20685) JOE-14-0356. Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, Zenere MB, Monauni T & Muggeo M 2000 Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin Declaration of interest sensitivity: studies in subjects with various degrees of glucose tolerance The authors declare that there is no conflict of interest that could be and insulin sensitivity. Diabetes Care 23 57–63. (doi:10.2337/diacare. perceived as prejudicing the impartiality of the research reported. 23.1.57) Chan SM, Sun R-Q, Zeng X-Y, Choong Z-H, Wang H, Watt MJ & Ye J-M 2013 Activation of PPARa ameliorates hepatic insulin resistance and steatosis in high fructose – fed mice despite increased endoplasmic Funding reticulum stress. Diabetes 62 2095–2105. (doi:10.2337/db12-1397) This work was supported by the National Program on Key Basic Research Chen H, Sullivan G & Quon MJ 2005 Assessing the predictive accuracy of Project of China (973 Program) (2012CB722401), the National Natural QUICKI as a surrogate index for insulin sensitivity using a calibration Science Foundation of China (81030051, 21177046, and 21207128), the R&D model. Diabetes 54 1914–1925. (doi:10.2337/diabetes.54.7.1914) Special Fund for Public Welfare Industry (Environment) (201309048), the Cohen JC, Horton JD & Hobbs HH 2011 Human fatty liver disease: old National Basic Research Development Program of China (2008CB418206), questions and new insights. Science 332 1519–1523. (doi:10.1126/ the Fundamental Research Funds for the Central Universities, Huazhong science.1204265) University of Science and Technology (2012QN240 and 2012TS072), and the Diamanti-Kandarakis E, Bourguignon J-P, Giudice LC, Hauser R, Prins GS, Natural Science Foundation of Fujian Province, China (2013J05033). Soto AM, Zoeller RT & Gore AC 2009 Endocrine-disrupting chemicals: an Endocrine Society scientific statement. Endocrine Reviews 30 293–342. (doi:10.1210/er.2009-0002) Donnelly KL, Smith CI, Schwarzenberg SJ, Jessurun J, Boldt MD & Parks EJ Author contribution statement 2005 Sources of fatty acids stored in liver and secreted via lipoproteins S X conceived and designed the study. J W and Y L performed the in patients with nonalcoholic fatty liver disease. Journal of Clinical researches and wrote the manuscript. X S carried out all western blot Investigation 115 1343–1351. (doi:10.1172/JCI23621) experiments and participated in data processing and statistical analyses. Fan JG, Zhu J, Li XJ, Chen L, Lu YS, Li L, Dai F, Li F & Chen SY 2005 Fatty Y L reviewed and edited the manuscript. L S, Z Z, and B X participated liver and the metabolic syndrome among Shanghai adults. Journal of

Gastroenterology and Hepatology 20 1825–1832. (doi:10.1111/j.1440- Reddy JK III 2001 Peroxisomal b-oxidation, PPARa, and steatohepatitis. 1746.2005.04058.x) American Journal of Physiology. Gastrointestinal and Liver Physiology 281 Farrell GC & Larter CZ 2006 Nonalcoholic fatty liver disease: from steatosis G1333–G1339. to cirrhosis. Hepatology 43 S99–S112. (doi:10.1002/hep.20973) Redgrave T 1970 Formation of cholesteryl ester-rich particulate lipid during Folch J, Lees M & Sloane-Stanley G 1957 A simple method for the isolation metabolism of chylomicrons. Journal of Clinical Investigation 49 465. and puriﬁcation of total lipids from animal tissues. Journal of Biological (doi:10.1172/JCI106255) Chemistry 226 497–509. Rubin BS & Soto AM 2009 Bisphenol A: perinatal exposure and body Friedman TC, Sinha-Hikim I, Parveen M, Najjar SM, Liu Y, Mangubat M, weight. Molecular and Cellular Endocrinology 304 55–62. (doi:10.1016/ Shin C-S, Lyzlov A, Ivey R & Shaheen M 2012 Additive effects of j.mce.2009.02.023) nicotine and high-fat diet on hepatic steatosis in male mice. Santra A, Chowdhury A, Ghatak S, Biswas A & Dhali GK 2007 Arsenic Endocrinology 153 5809–5820. (doi:10.1210/en.2012-1750) induces apoptosis in mouse liver is mitochondria dependent and is Hamaguchi M, Kojima T, Takeda N, Nakagawa T, Taniguchi H, Fujii K, abrogated by N-acetylcysteine. Toxicology and Applied Pharmacology 220 Omatsu T, Nakajima T, Sarui H & Shimazaki M 2005 The metabolic 146–155. (doi:10.1016/j.taap.2006.12.029) syndrome as a predictor of nonalcoholic fatty liver disease. Annals of Shankar A, Teppala S & Sabanayagam C 2012 Bisphenol A and peripheral Internal Medicine 143 722–728. (doi:10.7326/0003-4819-143-10- arterial disease: results from the NHANES. Environmental Health 200511150-00009) Perspectives 120 1297–1300. (doi:10.1289/ehp.1104114) Huc L, Lemarie´ A, Gue´raud F & He´lie`s-Toussaint C 2012 Low concen- Shen L, Liu Z, Gong J, Zhang L, Wang L, Magdalou J, Chen L & Wang H trations of bisphenol A induce lipid accumulation mediated by the 2014 Prenatal ethanol exposure programs an increased susceptibility production of reactive oxygen species in the mitochondria of HepG2 of non-alcoholic fatty liver disease in female adult offspring rats. cells. Toxicology In Vitro 26 709–717. (doi:10.1016/j.tiv.2012.03.017) Toxicology and Applied Pharmacology 274 263–273. (doi:10.1016/ Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G & j.taap.2013.11.009) Quon MJ 2000 Quantitative insulin sensitivity check index: a simple, Shimomura I, Matsuda M, Hammer RE, Bashmakov Y, Brown MS & accurate method for assessing insulin sensitivity in humans. Journal of Goldstein JL 2000 Decreased IRS-2 and increased SREBP-1c lead to Clinical Endocrinology and Metabolism 85 2402–2410. (doi:10.1210/ mixed insulin resistance and sensitivity in livers of lipodystrophic jcem.85.7.6661) and ob/ob mice. Molecular Cell 6 77–86. (doi:10.1016/S1097- Lang IA, Galloway TS, Scarlett A, Henley WE, Depledge M, Wallace RB & 2765(00)00009-5) Melzer D 2008 Association of urinary bisphenol A concentration with Singh S & Li SS 2012 Epigenetic effects of environmental chemicals medical disorders and laboratory abnormalities in adults. Journal of the bisphenol A and phthalates. International Journal of Molecular Sciences 13 American Medical Association 300 1303–1310. (doi:10.1001/jama.300. 10143–10153. (doi:10.3390/ijms130810143) 11.1303) Somm E, Schwitzgebel VM, Toulotte A, Cederroth CR, Combescure C, Nef S, Ma Y, Xia W, Wang D, Wan Y, Xu B, Chen X, Li Y & Xu S 2013 Hepatic DNA Aubert ML & Hu¨ppi PS 2009 Perinatal exposure to bisphenol A alters methylation modiﬁcations in early development of rats resulting from early adipogenesis in the rat. Environmental Health Perspectives 117 perinatal BPA exposure contribute to insulin resistance in adulthood. 1549–1555. (doi:10.1289/ehp.11342) Diabetologia 56 2059–2067. (doi:10.1007/s00125-013-2944-7) Tan H-H, Fiel MI, Sun Q, Guo J, Gordon RE, Chen L-C, Friedman SL, Marchesini G, Marzocchi R, Agostini F & Bugianesi E 2005 Nonalcoholic Odin JA & Allina J 2009 Kupffer cell activation by ambient air fatty liver disease and the metabolic syndrome. Current Opinion in particulate matter exposure may exacerbate non-alcoholic fatty liver Lipidology 16 421–427. (doi:10.1097/01.mol.0000174153.53683.f2) disease. Journal of Immunotoxicology 6 266–275. (doi:10.3109/ Journal of Endocrinology Marmugi A, Ducheix S, Lasserre F, Polizzi A, Paris A, Priymenko N, 15476910903241704) Bertrand-Michel J, Pineau T, Guillou H & Martin PG 2012 Low doses Tomaru M, Takano H, Inoue K-I, Yanagisawa R, Osakabe N, Yasuda A, of bisphenol A induce gene expression related to lipid synthesis and Shimada A, Kato Y & Uematsu H 2007 Pulmonary exposure to diesel trigger triglyceride accumulation in adult mouse liver. Hepatology 55 exhaust particles enhances fatty change of the liver in obese diabetic 395–407. (doi:10.1002/hep.24685) mice. International Journal of Molecular Medicine 19 17–22. (doi:10.3892/ Matthews D, Hosker J, Rudenski A, Naylor B, Treacher D & Turner R 1985 ijmm.19.1.17) Homeostasis model assessment: insulin resistance and b-cell function Tong M, Neusner A, Longato L, Lawton M, Wands JR & de la Monte SM from fasting plasma glucose and insulin concentrations in man. 2009 Nitrosamine exposure causes insulin resistance diseases: relevance Diabetologia 28 412–419. (doi:10.1007/BF00280883) to type 2 diabetes mellitus, non-alcoholic steatohepatitis, and McCurdy CE, Bishop JM, Williams SM, Grayson BE, Smith MS, Friedman JE Alzheimer’s disease. Journal of Alzheimer’s Disease 17 827–844. & Grove KL 2009 Maternal high-fat diet triggers lipotoxicity in the (doi:10.3233/JAD-2009-1155) fetal livers of nonhuman primates. Journal of Clinical Investigation 119 Utzschneider KM & Kahn SE 2006 Review: The role of insulin resistance in 323–335. (doi:10.1172/JCI32661) nonalcoholic fatty liver disease. Journal of Clinical Endocrinology and McGarry J & Foster D 1980 Regulation of hepatic fatty acid oxidation and Metabolism 91 4753–4761. (doi:10.1210/jc.2006-0587) ketone body production. Annual Review of Biochemistry 49 395–420. Veiga-Lopez A, Luense LJ, Christenson LK & Padmanabhan V 2013 (doi:10.1146/annurev.bi.49.070180.002143) Developmental programming: gestational bisphenol-A treatment Nagakawa Y, Williams GM, Zheng Q, Tsuchida A, Aoki T, Montgomery RA, alters trajectory of fetal ovarian gene expression. Endocrinology 154 Klein AS & Sun Z 2005 Oxidative mitochondrial DNA damage and 1873–1884. (doi:10.1210/en.2012-2129) deletion in hepatocytes of rejecting liver allografts in rats: role of TNF-a. Vom Saal FS, Nagel SC, Coe BL, Angle BM & Taylor JA 2012 The estrogenic Hepatology 42 208–215. (doi:10.1002/hep.20755) endocrine disrupting chemical bisphenol A (BPA) and obesity. Oben JA, Mouralidarane A, Samuelsson A-M, Matthews PJ, Morgan ML, Molecular and Cellular Endocrinology 354 74–84. (doi:10.1016/j.mce. Mckee C, Soeda J, Fernandez-Twinn DS, Martin-Gronert MS & 2012.01.001) Ozanne SE 2010 Maternal obesity during pregnancy and lactation Wei J, Lin Y, Li Y, Ying C, Chen J, Song L, Zhou Z, Lv Z, Xia W & Chen X programs the development of offspring non-alcoholic fatty liver 2011 Perinatal exposure to bisphenol A at reference dose predisposes disease in mice. Journal of Hepatology 52 913–920. (doi:10.1016/j.jhep. offspring to metabolic syndrome in adult rats on a high-fat diet. 2009.12.042) Endocrinology 152 3049–3061. (doi:10.1210/en.2011-0045)

Received in ﬁnal form 6 June 2014 Accepted 1 July 2014